Ask HN: How are you acquiring your first hundred users?

I have built this 3d bin packer. https://3dpack.ing Despite my best attempts didn’t take off:(

What did you try? It might be worth talking about here to try to gain some insight.

B2B products, unless miraculous, do not generally "take off" without a lot of hard work, meetings, trade shows, client demos, etc, at least so far as I have seen.

To make any sales at all, it seems like you'd have to already know a lot of people in the industry, and it would have to outcompete not only other products, but just the old "we'll load all the shit we can into the truck/container based on human intuition".

In fact, ideally, you'd ask someone you know (or even better, they'll ask you) if this is a problem you can solve for them. If you can't sell it to people you already know, it'll be even harder to sell to people you don't, presumably.

Plus it will be very hard to articulate to an AI "you can't sit that on top of that!", but it's obvious when looking at it.

I'm in the midst of wading into B2B, alone, and low on capital, myself. My "success" so far is mostly due to dumb luck and being available to talk to.

This looks really interesting! Optimisation problems like this are so much fun to solve and seem like such a natural fit for the SaaS model.

Could you give more of an idea about how much it "took off"? Based on earlier comments of yours, it seems that you have some paying customers. Has it grown very gradually? Do you have much churn?

Also curious how you have been splitting your time between algorithm development for the underlying optimisation problem (probably the most interesting part to me!), AI front-end (a neat value-add on top), marketing and dealing with support? If you have the time and energy, I'm sure a blog post or similar would be interesting to many people.

I guess i am around 150 MRR level or less perhaps. So it pays the costs and leaves me some money and I can run it forever given the condition satisfied.

I have spent countless months on the algorithm earlier. It improved from trash to something commercially reasonable gradually, it was more like getting stuck for a while then getting small break throughs.

My biggest mistake was changing my domain perhaps. Earlier it was bindrake.com Then I added 3d pack.ing and due to complexity of canonical pages google kicked me out.

Ironically recently I get contacted by a client and they said “you don’t exist on google but we found you from ChatGPT”

I had face to face meetings with some clients. Well to be fair 3d packing is also relatively competed area there are few popular products. Is my product better than them for some aspects yes. Hard to say what is wrong but perhaps me or the product isn’t seriously taken, yet it does add real value by solving a real problem by some means better than others.

Thanks! I know what you mean about "countless months" on an algorithm... Glad you were able to improve it to a useful level.

> it was more like getting stuck for a while then getting small break throughs

Like simulated annealing itself ;-) (Yes, I watched the deep dive clip)

Really cool that you give API examples right on the landing page BTW. Do you know if many customers use the API directly? Also really interesting to me that your algo is quick enough that you can return a solution directly before the HTTPS connection times out (I would have thought it might take half an hour to optimise, meaning you'd probably need to instead return some kind of handle that the user can use to pick up the solution via a different endpoint later).

A couple of minor nits I noticed, hope this is helpful rather than annoying:

- The "Blog" link at the top appears to produce a copy of the original landing page. Maybe just delete that link?

- The "API Response Structure" section is duplicated

- In the deep dive clip, at 5:31, the subtitle contains "three dpaq.ing", while at 5:54, both the voice and the subtitle say "third bin packer" (I suspect you meant "3D bin packer")

Yes, most of the project evolved because of potential customers who later disappeared. Someone shows up and says, “Hey, we’ll pay you if you provide this feature,” so I implement it… and then they vanish

The API came about the same way. Currently, there’s just one active API user: a company from Sri Lanka that built a Unity WASM interface on top of it.

I coded the algorithm completely from scratch, with no external inspiration and fun fact, it’s written in F# using functional programming. There are a lot of immutable lists floating around. Honestly, I probably wouldn’t have been able to write it without functional programming (even though it’s notoriously slow for this kind of problem). The immutability allowed me to focus on individual parts of the app without worrying about side effects.

That said, how it works and how fast it works still feels like magic to me, just like any meta-heuristic algorithm applied to combinatorial problems.

And thanks for the feedback! I knew the Blog link was broken after I redid the landing page, but I didn’t notice the duplicated API Response Structure section. I’ll definitely fix that. Though part of me thinks, “Should I really bother?”

Also, thanks for catching the audio/subtitle issues. I was aware of both, but I figured it wouldn’t make much difference at this stage. Maybe that attitude explains my low traction… who knows? Either way, I appreciate you taking the time to point these out!

Cool! And yes I'm a little surprised that you used a functional programming language for the implementation, I'll have to have a look at F#.

I think i fixed most issues, except the video one. Thanks again.